MSPA-CSE: Novel A Posteriori Analysis of Ecological Models: The Carbon Cycle

Colorado State University, Fort Collins CO

Investigators

Abstract

Many ecological models, such as the Integrated Regional Carbon Cycle Model describing the cycling of mass and energy and exchange of materials at the earth's surface, are characterized by complex, nonlinear interactions between different physical sub-systems acting on a vast range of scales. The investigators will develop and apply new tools for quantifying and controlling the effects of the uncertainty in a differential equation model, such as those that arise from experimental measurement, and for fitting such models to physical measurements. The new tools will be based on a combination of an a posteriori analysis technique that has been used with wide success in engineering and statistical methodology used for complex data sets. Models of the biosphere are becoming indispensable in the analysis and prediction of large-scale environmental change, including global warming. However, the complexity of modern models demands new approaches for analysis that combine deterministic and statistical mathematics in order to allow the rapid incorporation of new scientific understanding (e.g., new relevant processes) and data sources (e.g. new satellite technologies) and to quantify the effects of uncertainties in the representation of terrestrial systems. The combination of the power of the mathematical and statistical tools being employed in this project and the interdisciplinary mode of collaboration between the investigators yields substantial promise of significant advances on critically important issues in ecological modeling.